Contour analysis in problems of pattern recognition and objects classification

被引：0

作者：

Makarov, M.A. ^{[1
]}

Andreev, S.Y. ^{[1
]}

机构：

[1] Institute of Cybernetics, National Research Tomsk Polytechnic University, Tomsk, Russia

来源：

Scientific Visualization | 2014年 / 6卷 / 03期

关键词：

D O I：

暂无

中图分类号：

学科分类号：

摘要：

The article describes a method of describing the contour of the moving object using descriptors. Amplitude spectra of the Fourier transform was chosen as descriptors. Then etalon contours for comparison and classification were found and threshold conditions was defined of two types. Descriptor comparison was defined on two ways: compare with a correlation coefficient and compare with a lambda distances. © 2014, National Research Nuclear University. All rights reserved.

引用

页码：77 / 86